1- Using AI Writing Tools Ethically and Productively

Introduction: The Limitations of AI Writing Tools

ChatGPT and other AI writing assistants offer impressive capabilities but also have key limitations users should keep in mind:

• Hallucinations: ChatGPT can generate inaccurate or incoherent text since it is trained on a large dataset and may "hallucinate" text not based on real information.
• No web access: Without plugins, ChatGPT cannot access the most up-to-date information on the internet.
• Outdated information: All information from ChatGPT dates from before September 2020, so it may not be current or accurate.
• Incorrect sources: ChatGPT does not always provide correct sources, making it hard to verify accuracy.

The AI capabilities are remarkable, but the output cannot be assumed fully accurate or authoritative. Using AI thoughtfully and confirming key information is crucial. Tools like Perplexity.ai, for example, can help by reliably listing sources and generating useful follow-up prompts. It is worth experimenting with research-focused AI-powered tools, of which dozens of new ones are appearing every week.

Upholding Integrity with AI

In academia and professional fields, AI writing tools are now commonplace. Used appropriately, tools like ChatGPT help generate ideas, outline, and draft documents. However, users must uphold ethics and integrity. AI-generated content should not be passed off as original human writing. Proper attribution and citation of AI sources are still vital.

Generate by AI Bing Image creator, 24 October 2024

Here are examples of content requiring attribution:

• Summaries or topic overviews
• Lists of key ideas or bullet points
• Passages over a few sentences
• Data analysis or visualizations
• Creative writing like poems or stories
• Unique metaphors or phrases
• Market research reports
• Code or algorithms

If the AI significantly contributed work, creativity, analysis, or synthesis, it deserves attribution, even if sections are edited. Transparency regarding the AI's role is crucial for academic integrity.

Using AI Beyond Current Capabilities

A recent study of Boston Consulting Group consultants revealed professionals don't always know how to use AI correctly. For tasks within AI's capabilities, consultants were far more productive and effective using the technology. However, for complex problems beyond AI's current skills, consultants with AI were 19 percentage points less likely to find correct solutions. Developing abilities for responsible AI use is critical.

Building Skills for Responsible AI Use

Students today will encounter AI tools like ChatGPT in their careers and studies. A compelling vision in this sense was presented by Sam Kahn of the Kahn Academy in this Ted Talk. He maintains that Bloom's 2 sigma problem as outlined in his 1984 article, can finally overcome the inherent weakness of group, class-based teaching by offering super personal tutor for each student through a properly trained LLM.

Developing skills now for using these tools responsibly will be invaluable. Our approach is to always respect the guidelines and the rules of the school or university. In addition, working together transparently with teachers and lecturers will likely keep induce a positive learning process, and keep all parties out of trouble.

In our series on writing research essays, in the body of the essay with the data analysis and find, as well as the methodology section we suggest you write the first draft yourself without the use of any high-powered writing tools. Conversely, in the introduction, rationale, and conclusion sections you can use LLMs carefully to produce a first draft. More specialized research tools such as consensus, scite or elicit can be very helpful in identifying and analyzing secondary sources.

Focus on generating original ideas and connections with AI models, then produce your own valid work based on the inspiration. Evaluate critically whether to include any AI-generated text and cite it appropriately. These habits will lead to ethical and productive AI use that enhances rather than compromises your learning and work.

Conclusions

AI writing tools have limitations but also impressive capabilities. Users should be cautious in verifying accuracy. Upholding integrity by properly attributing AI work is vital. Developing abilities for responsible AI use focused on original thinking and appropriate citation will ensure AI acts as a beneficial tool rather than a crutch.

This blog post is followed by a series of other posts outlining how to develop the various sections of a research essay, and in what measure and how to use AI-powered tools for each section, sample prompts included:

1- Using AI Writing Tools Ethically and Productively

2- Developing a Clear and Focused Research Question and Title

3- Crafting a Strong Introduction for Your Research Essay

4- Research Methods Revealed: Writing the Methodology Section of Your Essay

5- Analyzing for Success: Creating a Strong Analysis Section

Works Cited

Dell'Acqua, F., McFowland III, E., Mollick, E., Lifshitz-Assaf, H., Kellogg, K. C., Rajendran, S., Krayer, L., Candelon, F., & Lakhani, K. R. (2023). Navigating the Jagged Technological Frontier: Field Experimental Evidence of the Effects of AI on Knowledge Worker Productivity and Quality. Working Paper. Harvard Business School. Retrieved from https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4573321